Analyzing Flight Data for an Emerging Hotelier
Project 3
Student 500599576
University of Sydney | DATA1001 | May 2021
Figure from marketwatch.com
1 Introduction
Traveling abroad has become more feasible compared to the past due to the growing Airline Industries. Nowadays, with the increasing number of flights to the country, the number of tourists has increased. This increase, in turn, has created a demand for hotels and hospitality industries.
1.1 Client
This report is aimed towards a potential entrepreneur who is looking forward to setting up his own hotel chain.The report pays particular attention to the cities with more flights, the airlines that operate in those areas, seasonal trends and some strategies that can be implemented to maximize the profit based on the trend in data.
2 Recommendaton
Cities such as Sydney, Melbourne, Brisbane, and Perth are ideal to set up the hotel. These cities are also suitable for extending the business.
Travelers from Asian countries have a strong influence in the tourism industry, thus top airlines or hotel chains can be valuable members when planning for business extension. Forming a partnership with the top airlines will be beneficial to capture more customers.
Summertime records a higher number of flights, there’s a greater chance of having more tourists. Hence, hotels have to take that into account and prepare accordingly.
2.1 Initial Look At Data
According to the data provided by the Bureau of Infrastructure and Transport Research Economics, the flight dataset contains 89312 observations of 15 variables. It reports the flights by airlines, route, city, country,etc., from 2003 to 2018.
# Reading the data
#International Airlines operating from Australia
flights = read.csv("flights.csv")
## Quick snapshot of the data
head(flights)3 Evidence and Interpretation
3.1 Which cities had the maximum number of flights?
The barplot represents the total number of flights from 2012 to different Australian cities, grouped by the months of a year. It is observed from the plot that the number of flights to Sydney is consistently higher than any other city. Flights to other cities such as Melbourne, Brisbane, and Perth follow right after. It appears that cumulatively the number of flights for these cities is higher in December, January, and March compared to other months of the year. The rise in the number of flights could be due to seasonal variation. So, it is deduced that the summer season draws more people to Australia as it’s the peak time for travelers to visit beaches and spend more time engaging in seasonal sporting events. (Budget Direct,2019)
library(tidyverse)
library(plotly)
# Selecting the inbound flights from 2012
graph1 = flights %>%
filter(Year >=2012 &
In_Out == "I") %>%
group_by(Month_num,
Year,
Australian_City) %>%
summarise(Sum_flights = sum(All_Flights)) %>%
ggplot(aes(x = Month_num,
y = Sum_flights,
fill = Australian_City)) +
geom_bar(stat = 'identity') +
scale_x_continuous(breaks = c(1,2,3,4,5,6,7,8,9,10,11,12)) +
labs( x = "\nMonths",
y = "Total Number of Flights\n",
title="Flights to the Australian Cities from 2012")
ggplotly(graph1, width = 750, height = 700)3.2 Do seasons play a role in the number of flights operating?
The plot represents how the total number of flights varied between summer and winter from 2010. It reveals the trend that flights to Australia has been dominant during the summer. The p-value of 0.0002379 also confirms that there is a seasonal variation and so summer is a crucial time for hoteliers as there are more festivals and occasions around this time to accommodate for. This means hotels need to carefully manage their accommodation, adjust their rates and pay attention to the staffing needs during this season.
# total flights during summer
summer_flights = flights %>%
filter(Month_num == 12 | Month_num == 1 | Month_num == 2 | Month_num ==3 & In_Out == "I") %>%
filter(Year >= 2010 & Year != 2018) %>%
group_by (Year) %>%
summarise(Sum_flights = sum(All_Flights)) %>%
rename(Sum_flights_summer = Sum_flights)
# total flights during winter
winter_flights = flights %>%
filter(Month_num == 5 | Month_num == 6 | Month_num == 7 | Month_num == 8 & In_Out == "I") %>%
filter(Year >= 2010 & Year != 2018) %>%
group_by(Year) %>%
summarise(Sum_flights = sum(All_Flights)) %>%
rename(Sum_flights_winter = Sum_flights)
seasonal_flights = data.frame(summer_flights, winter_flights$Sum_flights_winter)
ggplot(seasonal_flights,
aes(x = Year,
y = Sum_flights_summer)) +
geom_line(aes(y= Sum_flights_summer,
colour = "red"))+
geom_point(colour = "red",
size = 2)+
geom_line(aes(y =winter_flights.Sum_flights_winter,
colour= "blue"))+
geom_point(y = seasonal_flights$winter_flights.Sum_flights_winter,
colour ="darkblue", size = 2) +
labs(x = "Years",
y = "Total Number of Flights\n",
title = "Flights Coming to Australia") +
ylim(48000,60500) +
# (Muldoon, 2018) for creating the legend
scale_color_identity(name = "Seasons",
breaks = c("red", "blue"),
labels = c("Summer", "Winter"),
guide = "legend")+
scale_x_continuous(breaks = c(2010, 2011, 2012, 2013, 2014, 2015,2016,2017)) 
Hypothesis-
Null hypothesis = True mean is equal to 0
Alternate hypothesis = True mean is not equal to 0
Assumption - The population difference is normal and independent
Test statistic and P-value - The t value = 6.8689 and the p-value = 0.0002379
Since the p value is significant, null hypothesis is rejected.
total_flight_summer = c(summer_flights$Sum_flights_summer)
total_flight_winter = c(winter_flights$Sum_flights_winter)
flights_diff = total_flight_summer - total_flight_winter
t.test(flights_diff,mu = 0, var.equal = FALSE)##
## One Sample t-test
##
## data: flights_diff
## t = 6.8689, df = 7, p-value = 0.0002379
## alternative hypothesis: true mean is not equal to 0
## 95 percent confidence interval:
## 1303.222 2671.528
## sample estimates:
## mean of x
## 1987.3753.3 Which airlines had the most number of flights to these cities?
The Airlines that operate most of their flights to the cities mentioned above (Sydney, Melbourne, Brisbane) are Qantas Airways, Virgin Australia, and Jetstar. Given that the top 3 airlines with the maximum number of flights to those hotspots are Australian-based, it would be ideal to form a partnership with these airline companies. It will allow the access to more markets and networks that can help the hospitality business to grow. To maximize the profit and attract more people, advertisements can be directed towards these airlines. Many loyalty programs can also be introduced to reward long-time members leading to an increase customer satisfaction. This technique that has been carried out by Delta Airlines and Starwood Hotels that “has been wildly successful.” (Martin, 2020)
# Selecting the airlines that travel to the popular cities
airline = flights %>%
filter(In_Out == "I" & Year >= 2012 &
Australian_City == "Sydney"|
Australian_City == "Melbourne" |
Australian_City== "Brisbane") %>%
group_by(Airline) %>%
summarise(`Number of Flights` = n()) %>%
arrange(-(`Number of Flights`)) %>%
head(5)
graph2 = ggplot(airline,aes(x = Airline,
y = `Number of Flights` )) +
geom_bar(stat= 'identity', fill = "purple") +
labs(x = "Airlines",
title ="Airlines with the Maximum Number of Flights to the Top 3 Cities")
ggplotly(graph2)3.4 Which countries will be valuable as potential investors?
The graph shows the port countries that had over 12000 flights to the Australian cities from 2012. This depicts that Asian countries such as New Zealand, Singapore, China can play a massive role in boosting the Australian tourism industry. So it is recommended to consider airlines or other hospitality businesses from these countries for further growth of the industry. This is also consistent with the existing reports where Singapore has been a long-time partner in Australian Tourism. (Singapore – long-term partner in Australian tourism. (n.d.))
#(Sauer, 2016) for resizing the figure
flights %>%
filter(Year >= "2012" & In_Out == "I") %>%
group_by(Port_Country, Australian_City) %>%
summarise(sum_flights = sum(All_Flights)) %>%
filter(sum_flights >12000) %>%
ggplot(aes(x = Port_Country,
y = sum_flights,
fill = Australian_City)) +
geom_bar(stat = "identity")+
theme(axis.text.x = element_text(angle = 90)) +
labs(x= "Port Countries",
y = "Total Flights ",
title = "Countries with Over 12000 Flights to the Australian Cities")
4 Other Evidence
Normality of the data can be safely assumed, of course, having more data would have given a more concrete proof.
## QQ plot
library(ggplot2)
qqnorm(flights_diff, col= "black")
qqline(flights_diff, col = "blue")
There is a significant amount of variation, so the t.test() is set to var.equal = FALSE
# Boxplot
boxplot(total_flight_summer,total_flight_winter, names = c("Summer Flights", "Winter Flights"), col = "lightpink")
5 References
Budget Direct. (2019, September 1). Australian Tourism Industry 2020: Travel Research & Statistics. Budget Direct. https://www.budgetdirect.com.au/travel-insurance/research/tourism-statistics.html.
Weather in Australia. Tourism Australia. (n.d.). https://www.australia.com/en/facts-and-planning/weather-in-australia.html.
Martin, G. (2020, December 15). Airlines And Hotels Team Up To Reward Each Other’s Loyalty Program Members. Forbes. https://www.forbes.com/sites/grantmartin/2014/09/30/airlines-and-hotels-team-up-to-reward-each-others-loyalty-program-members/?sh=faad47918fd3.
Singapore – long term partner in Australian tourism. (n.d.). https://www.tourisminvestment.com.au/en/research-insights/news/Singapore-Long-Term-Partner-in-Australian-Tourism.html.
Sauer, S. (2016, November 2). Different ways to set figure size in RMarkdown. Sebastian Sauer Stats Blog. https://sebastiansauer.github.io/figure_sizing_knitr/.
Muldoon, A. (2018, July 19). Creating legends when aesthetics are constants in ggplot2. Very statisticious. https://aosmith.rbind.io/2018/07/19/manual-legends-ggplot2/.